No abstract

By in situ hybridization, 44-100% of the blood eosinophils from five patients with hypereosinophilia and four normal subjects exhibited intense hybridization signals for TNF-a mRNA. TNF-a protein was detectable by immunohistochemistry in blood eosinophils of hypereosinophilic subjects, and purified blood eosinophils from three atopic donors exhibited cycloheximide-inhibitable spontaneous release of TNF-a in vitro. Many blood eosinophils (39-91%) from hypereosinophilic donors exhibited intense labeling for macrophage inflammatory proteinla (MIP-la) mRNA, whereas eosinophils of normal donors demonstrated only weak or undetectable hybridization signals for MIP-la mRNA. Most tissue eosinophils infiltrating nasal polyps were strongly positive for both TNF-a and MIP-la mRNA. By Northern blot analysis, highly enriched blood eosinophils from a patient with the idiopathic hypereosinophilic syndrome exhibited differential expression of TNF-a and MIP-la mRNA. These findings indicate that human eosinophils represent a potential source of TNF-a and MIP-la, that levels of expression of mRNA for both cytokines are high in the blood eosinophils of hypereosinophilic donors and in eosinophils infiltrating nasal polyps, that the eosinophils of normal subjects express higher levels of TNF-a than MIP-la mRNA, and that eosinophils purified from the blood of atopic donors can release TNF-a in vitro. (J. Clin. Invest. 1993. 91:2673-2684

SummaryTransforming growth factor a (TGF-a) is a pleuripotential cytokine with diverse biological effects, including the ability to influence the proliferation of normal cells or neoplastic epithelial cells. Eosinophils are a subset of granulocytes that normally enter the peripheral tissues, particularly those beneath gastrointestinal, respiratory, and urogenital epithelium, where they reside in close proximity to the epithelial elements . In this study, we demonstrate that the great majority of eosinophils infiltrating the interstitial tissues adjacent to two colonic adenocarcinomas and two oral squamous cell carcinomas labeled specifically by in situ hybridization with a 35S-riboprobe for human TGF-a (hTGF-a) . No other identifiable leukocytes in these lesions contained detectable hTGF-a mRNA. We also examined leukocytes purified from a patient with the idiopathic hypereosinophilic syndrome. 80% of these eosinophils, but none of the patient's neutrophils or mononuclear cells, were positive for hTGF-a mRNA by in situ hybridization, and 55% of these eosinophils were positive by immunohistochemistry with a monoclonal antibody directed against the COOH terminus of the mature hTGF-a peptide. Finally, the identification of the purified eosinophil-associated transcript as hTGF-ca was confirmed by polymerase chain reaction product restriction enzyme analysis followed by Southern blot hybridization . In contrast to eosinophils from the patient with hypereosinophilic syndrome, the peripheral blood eosinophils from only two of seven normal donors had detectable TGF-a mRNA and none ofthese eosinophils contained immunohistochemically detectable TGF-a product. Taken together, these findings establish that human eosinophils can express TGF-a, but suggest that the expression of TGF-a by eosinophils may be under microenvironmental regulation. Demonstration of TGF-a production by tissueinfiltrating eosinophils and the eosinophils in the hypereosinophilic syndrome identifies a novel mechanism by which eosinophils might contribute to physiological, immunological, and pathological responses .

Aberrant expression of TGF-alpha is associated with human malignant oral epithelium. Experiments were initiated to determine the cellular sources of transforming growth factor-alpha (TGF-alpha) in human oral cancer. Ten freshly resected human oral cancers and four specimens of normal human oral epithelium were studied by in situ hybridization and immunohistochemistry. Tissues were probed with 35S-labeled sense and antisense riboprobes to (i) human TGF-alpha (hTGF-alpha), (ii) human epidermal growth factor receptor (EGFR) to determine the distribution of TGF-alpha responsive cells, and (iii) histone H3 to examine TGF-alpha and/or EGFR's possible contribution to altered proliferation in transformed epithelium. Results of our experiments showed that TGF-alpha mRNA could be detected in normal and transformed human oral epithelium. More surprising, we have identified the major source of TGF-alpha mRNA to be the infiltrating eosinophils. A monoclonal antibody to the mature human TGF-alpha peptide stained similar areas in normal and malignant specimens. Eosinophils associated with tumors exhibited positive cytoplasmic immunostaining for TGF-alpha protein. Labeling of EGFR mRNA in human oral epithelium demonstrated uniform labeling of basal layers in normal, hyperplastic, and mildly dysplastic epithelium. In severely dysplastic epithelium and carcinomas (particularly moderate to poorly differentiated types), cellular levels of EGFR mRNA were significantly higher. The profile of altered cellular levels of EGFR mRNA correlated well with the profile of altered proliferation as indicated by H3 mRNA labeling. We hypothesize that the overproduction of EGFR mRNA in tumor epithelium--together with the localized delivery of high amounts of TGF-alpha by eosinophils at tumor-developing sites--is responsible for the increased proliferation of the tumor epithelium.